AI-assisted English-as-an-Additional Language Benchmark Assessment Application

The Vendor is required to provide AI-assisted English-as-an-additional language benchmark assessment application for include:
• Discovery activities to validate business requirements, technical feasibility, AI capabilities, privacy considerations, implementation sequencing, user experience, and solution architecture prior to completion of detailed solution design and development.
• Configure and implement the solution using province-provided benchmark criteria, benchmark indicators, assessment rubrics, writing exemplars, assessment samples, approved training datasets, and other educational artefacts supplied by the province.
• AI-assisted English as an additional language (EAL) benchmark assessment application that supports kindergarten to grade 6 students and aligns with the state EAL benchmark framework.
• Collaborate with the province, designated EAL specialists, educators, and other subject matter experts throughout discovery, design, development, testing, and implementation to validate business requirements, benchmark criteria, AI outputs, and solution functionality.
• Configurable solution capable of assessing English language proficiency across the four language strands (listening, speaking, reading, and writing), recognizing that implementation of individual assessment capabilities may be sequenced based on technical feasibility, the phased implementation
• AI-assisted assessment capabilities for typed, handwritten, image-based, and spoken student submissions that generate explainable benchmark recommendations supported by transparent assessment rationale using a hybrid speech architecture that separates speech transcription (utilizing an open-source multimodal model such as phi-4-multimodal running inside the sovereign compute environment) from a dedicated, non-native acoustic phonetic and fluency scoring engine.
• Solution that preserves teacher professional judgement by requiring teacher review and validation of all AI-generated benchmark recommendations and providing the ability to override recommendations with documented rationale.
• Governance capabilities that support province-controlled review, approval, management, versioning, and auditability of benchmark criteria, approved training datasets, teacher feedback, benchmark overrides, prompts, and AI model refinement activities.
• Configurable AI model governance capabilities that enable the province to approve, manage, monitor, validate, version, retrain, retire, and audit AI models, benchmark criteria, prompts, approved training datasets, and associated model artefacts throughout the solution lifecycle.
• Deliver, operate, maintain, and support the solution such that all application components, AI models, processing services, storage, logging, monitoring, backup, recovery, and supporting services remain within the approved sovereign compute environment throughout the contract term.